Electronic check processing and in particular electronic check image processing has provided for many increased efficiencies in check writing and processing. Devices are able to image checks, check images can be electronically read, and the read data and images can be submitted for processing. However, various algorithms for reading text present on checks can be error prone, even when the particular algorithm is considered quite reliable. Poorly read data can present issues such as processing checks for amounts greater than what they are written for, processing checks with incorrect dates, drawing checks against incorrect accounts, and the like. To avoid such errors, it is common for involved systems to request manual verification of data. However, this increases the expense of such solutions and defeats many of the promises of electronic check processing.